On Tue, Jun 7, 2011 at 1:50 PM, H.J. Lu wrote:
> On Wed, Jun 1, 2011 at 2:23 AM, Ira Rosen wrote:
>> Hi,
>>
>> The vectorizer expects widening multiplication pattern to be:
>>
>> type a_t, b_t;
>> TYPE a_T, b_T, prod_T;
>>
>> a_T = (TYPE) a_t;
>> b_T = (TYPE) b_t;
>> prod_T =
On Wed, Jun 1, 2011 at 2:23 AM, Ira Rosen wrote:
> Hi,
>
> The vectorizer expects widening multiplication pattern to be:
>
> type a_t, b_t;
> TYPE a_T, b_T, prod_T;
>
> a_T = (TYPE) a_t;
> b_T = (TYPE) b_t;
> prod_T = a_T * b_T;
>
> where type 'TYPE' is double the size of type
On Mon, Jun 6, 2011 at 3:04 PM, Richard Sandiford
wrote:
> Richard Guenther writes:
>> Thanks. I would hope that we eventually can get rid of the
>> pattern recognizer ... at least for SSE there is also always
>> a scalar variant instruction for each vectorized one.
>
> AFAIK, that isn't true fo
Richard Guenther writes:
> Thanks. I would hope that we eventually can get rid of the
> pattern recognizer ... at least for SSE there is also always
> a scalar variant instruction for each vectorized one.
AFAIK, that isn't true for ARM and NEON. E.g. I don't know of a single
instruction that do
On Thu, Jun 2, 2011 at 1:08 PM, Ira Rosen wrote:
> On 2 June 2011 12:59, Richard Guenther wrote:
>> On Thu, Jun 2, 2011 at 10:46 AM, Ira Rosen wrote:
>>> On 1 June 2011 15:14, Richard Guenther wrote:
On Wed, Jun 1, 2011 at 1:37 PM, Ira Rosen wrote:
> On 1 June 2011 12:42, Richard Guen
On 2 June 2011 12:59, Richard Guenther wrote:
> On Thu, Jun 2, 2011 at 10:46 AM, Ira Rosen wrote:
>> On 1 June 2011 15:14, Richard Guenther wrote:
>>> On Wed, Jun 1, 2011 at 1:37 PM, Ira Rosen wrote:
On 1 June 2011 12:42, Richard Guenther wrote:
> Did you think about moving pass_
On Thu, Jun 2, 2011 at 10:46 AM, Ira Rosen wrote:
> On 1 June 2011 15:14, Richard Guenther wrote:
>> On Wed, Jun 1, 2011 at 1:37 PM, Ira Rosen wrote:
>>> On 1 June 2011 12:42, Richard Guenther wrote:
>>>
Did you think about moving pass_optimize_widening_mul before
loop optimizations?
On 1 June 2011 15:14, Richard Guenther wrote:
> On Wed, Jun 1, 2011 at 1:37 PM, Ira Rosen wrote:
>> On 1 June 2011 12:42, Richard Guenther wrote:
>>
>>> Did you think about moving pass_optimize_widening_mul before
>>> loop optimizations? Does that pass catch the cases you are
>>> teaching the p
On Wed, Jun 1, 2011 at 1:37 PM, Ira Rosen wrote:
> On 1 June 2011 12:42, Richard Guenther wrote:
>
>> Did you think about moving pass_optimize_widening_mul before
>> loop optimizations? Does that pass catch the cases you are
>> teaching the pattern recognizer? I think we should try to expose
>>
On 1 June 2011 12:42, Richard Guenther wrote:
> Did you think about moving pass_optimize_widening_mul before
> loop optimizations? Does that pass catch the cases you are
> teaching the pattern recognizer? I think we should try to expose
> these more complicated instructions to loop optimizers.
On Wed, Jun 1, 2011 at 11:23 AM, Ira Rosen wrote:
> Hi,
>
> The vectorizer expects widening multiplication pattern to be:
>
> type a_t, b_t;
> TYPE a_T, b_T, prod_T;
>
> a_T = (TYPE) a_t;
> b_T = (TYPE) b_t;
> prod_T = a_T * b_T;
>
> where type 'TYPE' is double the size of type
Hi,
The vectorizer expects widening multiplication pattern to be:
type a_t, b_t;
TYPE a_T, b_T, prod_T;
a_T = (TYPE) a_t;
b_T = (TYPE) b_t;
prod_T = a_T * b_T;
where type 'TYPE' is double the size of type 'type'. This works fine
when the types are signed. For the unsign
12 matches
Mail list logo
